How does Cep57 organize pericentriolar material to prevent mosaic variegated aneuploidy?

Cep57 tethers the Cep63 platform to centriolar microtubules

The paper's answer to how Cep57 prevents mosaic variegated aneuploidy (MVA) is architectural rather than enzymatic. Cep57 holds two structurally different connections at once: an oligomeric C-terminal region that binds centriolar microtubules, and a dimeric N-terminal region that engages Cep63 through a conserved bundle-like interface. Because Cep63 and Cep152 form the platform that positions the kinase Plk4, this bridge effectively fastens the machinery that starts a new centriole onto the existing centriole's microtubule wall.

That framing explains why two very different CEP57 mutations can produce the same MVA2 syndrome. C-terminal truncating mutations remove the microtubule-binding end, while the N-terminal exon 3 deletion removes the Cep63-binding end; the paper argues that losing either end breaks the same physical link. The rest of the study tests that claim at the level of nanoscale position, purified protein structure, and patient-derived cells.

MINFLUX places Cep57 beside centriolar microtubules

To see where Cep57 actually sits in the crowded pericentriolar space, the authors used 3D MINFLUX nanoscopy, a localization method that maps individual fluorescent emitters with nanometer-scale precision. Cep57 localizations from immunostained U2OS cells were filtered with a 3-nm interposition cutoff, clustered with DBSCAN, and fitted to a nine-fold symmetric cylindrical template. The combined map showed elongated Cep57 zones about 200 nm high arranged in nine-fold symmetry, with heterogeneous clustering.

A parallel MINFLUX map of acetylated tubulin marked the mother centriole's microtubules at roughly 500 nm in height. Because two-color MINFLUX was technically difficult, the authors merged the two density maps by concentrically aligning their nine axes and leveling their proximal ends. In that merged view, a substantial fraction of Cep57 and acetylated tubulin localizations lay close together, and Cep57 signal was confined to a 5 aL volume of the proximal pericentriolar space. This spatial coincidence is consistent with, but does not by itself prove, microtubule-directed localization; the direct binding evidence comes from purified proteins.

Only Cep57, not its paralog Cep57L1, bundles microtubules

The paralog comparison gives the mechanism a sharp edge. Cep57 and Cep57L1 share sequence homology and were previously reported to act redundantly in centriole duplication, yet only CEP57 mutations have been linked to MVA. The authors found a conserved α-helix (residues 301–316) in Cep57 that is absent or functionally distinct in Cep57L1, and the MVA2-linked p.L309Pfs*9 truncation disrupts it.

In maltose-binding protein pull-downs, Cep57-C (residues 272–500) bound both unmodified and acetylated tubulin, while Cep57L1-C did not. In microtubule cosedimentation at 2.5 μM protein, a concentration comparable to endogenous Cep57, Cep57-C coprecipitated with Taxol-stabilized microtubules and Cep57L1-C did not. Cep57-C also strongly enhanced microtubule polymerization, directly bound in vitro-generated microtubules, and stably bundled them; negative-stain electron microscopy showed protein aggregates along the polymerized microtubules with Cep57-C but not Cep57L1-C. These are in vitro properties of purified domains, so they establish biochemical capability rather than the cellular stoichiometry.

Oligomerization is what lets Cep57 anchor and function

Size-exclusion chromatography showed Cep57-C eluting mainly as a high-molecular-weight oligomer with a minor dimeric fraction, whereas Cep57L1-C eluted mainly as a dimer or low-molecular-weight complex. Analytical ultracentrifugation agreed: Cep57-C was polydisperse with most material sedimenting between 10 and 20 S, while Cep57L1-C was mostly a 3.73 S monomer (43%) with only minor faster species. The 277–334 region containing the conserved α-helix was sufficient to form higher-molecular-weight oligomers.

This matters because the paper links oligomerization to anchoring. Disrupting the α-helix (301–316) motif deprived Cep57 of its ability to oligomerize, stably bind centriolar microtubules, and promote centriole duplication. The authors interpret oligomerization as a way to reach high local concentration in the pericentriolar space, and they note that Cep57 has been reported to undergo liquid–liquid phase separation. That interpretation is a proposed rationale, not a directly measured consequence in this study.

The Cep57–Cep63 interface is a 2:2 helical bundle

On the other side of the bridge, the authors solved crystal structures of the Cep57(65–115) dimer (PDB 21JQ) and the Cep57(63–120)•Cep63(1–87) complex (PDB 9I7Q). The complex is a 2:2 heterotetrameric α-helical bundle. Mutating the hydrophobic core to alanine (4A) or to lysine (2K) in either partner greatly impaired the Cep57–Cep63 interaction, while mutations aimed at the Cep63 N-terminal 1–11 region (2D and Δ11) only weakly affected it.

The complex lacking Cep63 residues 1–11 sedimented largely as a 2.30 S species but also formed concentration-dependent large aggregates of 4–6 S, where an octamer would be expected near 3.6 S. The authors conclude that the Cep57•Cep63 complex functions primarily as a heterotetrameric bundle, with the Cep63 N-terminal region playing an auxiliary role in tuning the equilibrium toward higher-order assemblies. This is a structural and biochemical conclusion about purified components; it does not measure the assembly state inside cells.

Breaking either side of the bridge blocks centriole duplication

Cellular tests asked whether the interface matters for function. In U2OS cells expressing wild-type or mutant Cep57 while endogenous Cep57 and Cep57L1 were silenced, the 4A and 2K mutants were expressed comparably to wild type but showed greatly diminished centrosome-localized signal and significantly reduced Cep63 intensity. They also impaired normal duplication of centrosome-localized SAS6 and CP110 and increased the population with ≥3 SAS6 and ≥5 CP110 signals, a pattern previously attributed to precocious centriole disengagement.

The reciprocal experiment gave the same picture. Under Cep63 silencing, Cep63 4A and 2K mutants defective in Cep57 binding failed to localize themselves and Cep57 to centrosomes and failed to properly recruit SAS6 and CP110. The 2D and Δ11 mutants, which only moderately weakened the interaction, showed only a low level of defect. Together these experiments show that the two proteins stabilize each other's centrosomal localization and that the bundle interface is required for normal centriole duplication, while also allowing a minor amplification phenotype when it is disrupted.

Patient T cells show duplication failure as the dominant defect

The clinically decisive evidence comes from primary T lymphocytes purified from peripheral blood of affected siblings diagnosed with MVA2, their heterozygous father, and an unrelated healthy control. The affected cells carry the Cep63-binding-defective CEP57^Δexon3 mutation. In these patient cells, the predominant phenotype was a centriole-duplication defect, accompanied by a minor centriole-amplification component.

The authors interpret the amplification as likely resulting from disorganized PCM-induced premature centriole disengagement. That wording matters: the duplication failure is the measured dominant outcome, while the causal link from disorganized PCM to premature disengagement is the authors' explanation. The sample is small and drawn from one family's mutation, so the result establishes a mechanism for this MVA2 genotype rather than a general rule for all CEP57 mutations or all MVA cases.

Why the dual-module model changes the reading of MVA2

The paper's central priority is that Cep57 should be read as a two-ended architectural organizer, not as a single binding activity. The MINFLUX map, the purified-domain biochemistry, the crystal structures, and the patient-cell phenotypes converge on the same point: the Cep63-binding end and the microtubule-binding end are both required, and losing either produces MVA2. This also explains why Cep57 and Cep57L1 are not interchangeable despite their homology, since the Cep57 C-terminal α-helix supports oligomerization and microtubule bundling that Cep57L1-C did not show in these assays.

For readers weighing the evidence, the most useful split is between what was measured and what was inferred. Microtubule binding, bundling, oligomerization, the 2:2 bundle structure, and the mutant phenotypes are direct results. The claim that disorganized PCM causes premature disengagement in patient cells, and the phase-separation rationale for oligomerization, are interpretive extensions. The practical threshold the paper sets is therefore specific: a CEP57 variant that preserves only one of the two architectural modules should be treated as potentially pathogenic for MVA2, and the duplication defect, not the minor amplification, is the phenotype to expect first.

Sources used in this answer